Web-based tool for searching tables’ contents

• Autor(a) principal: Oliveira, Alexandre Daniel Moreira
• Data de Publicação: 2018
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/25876
• topic: Text Mining; Table Mining; Concept Recognition; Information Retrieval; Bioinformatics

The number of biomedical articles is constantly growing and researchers have more and more difficulty to efficiently find relevant information, compare results and identify new hypotheses. Text mining techniques have been explored to develop systems with the aim of providing easy and fast access to scientific literature. The problem is that most of these tools completely ignore tables and just process textual parts. This dissertation focuses on the analysis and indexing of tables extracted from scientific articles, as they often include a lot of information that can be useful to researchers and it is not available in the remaining content of the publications. So, the main objective of the work is to create a flexible indexing structure to handle different table formats and recognize biomedical concepts referred in the tables themselves, their captions and texts that reference them. A web-based tool was developed to allow users to search and visualize annotated tables extracted from scientific articles. The solution found uses some open-source frameworks, namely Neji for concept recognition and Elasticsearch for text indexing.